The purpose of the research is to develop new mathematical models and new algorithms for intelligent agents with neural network controllers in maritime transport. Main goal of research is increasing of safety level of maritime transport. The research includes the analysis and modelling of intelligent electronics devices based on multi-agent approaches. Methodologies used in research are intelligent electronic devices and negotiations, artificial neural network controllers, bond graphing for mechatronic system analysis and control. The research proposes the solution for control task between intelligent maritime transport units and intelligent maritime control units with a purpose to prevent accidents. The model includes algorithms of intelligent ship mechatronic system control and intelligent schedule control.